Finance Pulse

What Finance Pulse MCP server does

Finance Pulse MCP server connects MCP clients to the Finance Pulse API, which represents financial news as structured statements. Each statement can include a one-sentence fact, sentiment, importance, tickers, sector, source, headline, and publication time. Statements about the same development are grouped into topics; related topics about one company can form a storyline, while recurring subjects are represented as themes.

The server fits the Finance & Fintech category and is intended for assistants that need to retrieve, filter, and summarize financial news rather than browse article pages directly. Results include source information and can be queried by company, sector, theme, topic, storyline, sentiment, importance, and time range.

Tools and capabilities

Finance Pulse MCP server provides tools for several levels of news analysis:

• Resolve a company name to its symbol with find_symbol.
• Retrieve a company or other entity overview with get_symbol, including period counts, sentiment, major topics, storylines, related entities, and recent statements.
• Filter individual statements with search_statements.
• Find current or recently prominent developments with trending and find_topics.
• Inspect one topic with get_topic, including its statement mix, storylines, and latest outlets.
• Group a company’s coverage into continuing stories with storylines, then inspect one with get_storyline.
• Request daily or hourly volume and sentiment through sentiment_series.
• Rank symbols or sectors by news attention and change with screen.
• Explore recurring subjects with themes.
• Check available data, theme identifiers, and accepted filters with reference.

Each tool call normally maps to one API request. Responses are returned in small pages and can include a continuation cursor.

Setup and configuration

Finance Pulse MCP server runs through uvx, which is provided by uv; the package does not require a separate installation for MCP use. Configure the RAPIDAPI_KEY environment variable with a key from a RapidAPI account subscribed to Finance Pulse. The Basic plan is described as free and sufficient for trying the available functionality, but an unsubscribed key receives a 403 response.

Claude Code can add the server with the claude mcp add command using stdio transport. Claude Desktop and Cursor use an MCP configuration containing uvx as the command, finance-pulse as its argument, and RAPIDAPI_KEY in the environment. On Claude Desktop, restart the application after changing its configuration. If Claude Desktop cannot locate uvx, use the executable’s full path.

How it works

The underlying API supplies a data window and timing information in its snapshot, including the window start, anchor, and possible lag. Statements also carry publication and indexing timestamps. The API supports paged retrieval; the Python package documentation notes a default stopping point for iterator methods and a page size of 100 for larger statement reads, although the MCP tools expose the query functions rather than the Python iteration interface.

The data model supports both point-in-time investigation and time-series analysis. For example, an assistant can identify a ticker, inspect its current storylines, retrieve the newest sourced statements, and then request daily sentiment for a selected period. screen can compare attention across symbols or sectors, while trending surfaces developments reported by the most outlets.

Limitations and notes

Access depends on the subscribed RapidAPI plan’s quota and rate limits. A 429 response indicates that the quota or rate limit has been exhausted. Network failures, timeouts, and server errors can also affect requests. The available results cover the API’s data window rather than an unlimited historical archive; use reference to inspect the current window and supported filter values. The server retrieves structured data from the API, so it should not be treated as a replacement for reading full source articles.